How to Optimize Prices Using AI

One of the reasons why retailers increasingly rely on AI-based Dynamic Pricing is the huge amount of data and influencing factors that need to be considered for price determination, including product master data, purchase prices, sales quotas, inventory levels, competitor behavior, promotions, real-time transactions, historical data, seasonal trends, regional factors, weather influences - to name a few.

The crucial point is: the amount of data available is huge. Huge, and full of potential - especially in the age of AI, as an artificial system is easily able to process large amounts of data in seconds and derive meaningful analyses. And when it comes to dynamic price optimization, outcomes are evaluated numerically, which makes it especially easy to determine ROI and optimize strategies moving forward…

So, what do you need to use an AI for price optimization? It’s quite simple:

  • Sales and transaction data
  • Product master data
  • 4 to 6 weeks of history for ecommerce, 9 to 12 weeks for physical stores

Sales and transaction data

AI requires sales and transaction data in order to calculate the demand for each product. Price elasticity forms the basis for every price decision of the AI. Adding in transaction and shopper interaction data. improves the AI's forecasting quality and enhances its results – for example, viewed products, created shopping baskets, cancelled shopping baskets, saved watch lists or entered search terms. It makes sense to ensure this data is continuously available - either via real-time tracking like dynamic pricing software, or through a data feed from a SAP CAR system. 

Product master data

Product master data is the digital representation of your assortment, and as a result, is an important tool for dynamic price optimization. Master data provides information such as product ID, master-variant assignment, current price, RRP, lower and upper price limit, seasonal identification, brand, color, size, stock level, expiration date or target sales date. An AI can use these attributes for various tasks, including:

  • Combining existing supply with demand, allowing AI to calculate optimal prices in line with the market.
  • Using the best-before or target sales dates for to recognize when perishable items are going out of stock and pricing them according to their demand in order to reach zero stock on the given target date while protecting margin.
  • Identifying similarities between products with attributes such as color, brand and size. AI is able to calculate optimal prices for products with low data availability, long-tail items or even new products.
  • Optimizing price within a pre-determined range of price floors and ceilings.
  • Implementing family pricing and/or maps product relations, such as the appropriate price range between low-budget and branded products, by means of the master-variant assignment.

Implementation time

The implementation of a price optimization software is carried out in three steps:

1. The software is delivered as a cloud service.
2. The automated delivery of input data is set up.
3. The AI is configured and set to your objectives via a Web GUI.

Afterwards, the system goes live and continuously optimizes prices. For projects in e-commerce, this often takes no longer than 4 to 6 weeks, since many processes - by their nature - already run digitally. For physical retail it usually takes 9 to 12 weeks, but this up-front investment is essential for long-term results.

Now only one question remains: When will you go live?